For details about how to handle the extraction system definition file, see 5.2.2 Modifying defined information.

For details about how to handle the extraction environment definition file, see 5.3.2 Modifying defined information.

For details about how to handle the transmission environment definition file, see 5.4.2 Modifying defined information.

For details about how to handle the extraction definition file, see 5.5.2 Modifying defined information.

The following explains how to handle the extraction definition preprocessing file.

Converting the extraction definition file again

If you have modified definitions for a table subject to extraction processing at the source HiRDB or you have modified the source Datareplicator's extraction definition, you must convert the extraction definition file to the internal format again.

To convert the extraction definition file to the internal format:

1. Terminate the source Datareplicator normally.
2. Modify the definitions for the table subject to extraction processing or the extraction definition.
3. At the source system, execute the hdeprep command.
4. Start the source Datareplicator normally.

The following explains how to handle the extraction information queue files. For details about the handling procedure when an extraction information queue file is full, see 9.1.2 Error handling methods.

(a) Modifying a filename, the file size, or the number of files

When using a UNIX regular file or Windows file:

1. Terminate the source system.
2. Terminate the source Datareplicator normally.
3. Use a text editor to modify the corresponding operands in the extraction environment definition.
4. Execute initial start on the source Datareplicator.
5. Start the source Datareplicator normally.
6. Start the source system.

When using a UNIX character special file:

1. Terminate the source system.
2. Terminate the source Datareplicator normally.
3. Use an OS command to re-create the extraction information queue file in character special file format.
   

   Note

   Perform this step only when a status file size or the disk arrangement has changed for a reason such as the addition or removal of HiRDB servers.

4. Use a text editor to modify the corresponding operands in the extraction environment definition as appropriate for the created extraction information queue file.
   

   Note

   Perform this step only when a status file size or the disk arrangement has changed for a reason such as the addition or removal of HiRDB servers.

5. Execute initial start on the source Datareplicator.
6. Start the source Datareplicator normally.
7. Start the source system.

(b) Command for changing the organization of extraction information queue files (hdemodq command)

The hdemodq command enables you to change the organization of the extraction information queue files without having to initialize the source Datareplicator. The following table describes the operations supported by the hdemodq command:

hdemodq command operation	Description
Displaying information about the extraction information queue files	You can display information about the extraction information queue files in offline mode.
Registering additional extraction information queue files	To avoid a file full status, you can use this command to add extraction information queue files.
Releasing registered extraction information queue files	You can delete any extraction information queue file that is no longer needed.

Prerequisites for executing the hdemodq command

You can execute the hdemodq command under the following conditions:

• The source Datareplicator is inactive.
• The HDEPATH environment variable has been set.
• The extraction information queue file whose organization is to be changed is located at the node where the command is executed.
• The user executing this command also executes the extraction node master process in a UNIX system (specified in inetd.conf) or has the Administrators privilege in a Windows system. General users can execute this command only when the -l option is specified in the hdemodq command (to display file information).

You can execute multiple hdemodq commands concurrently at the same source system only to display information. Such concurrent command execution is not supported for registration of additional files or release of file registration (if such an attempt is made, all but one of the commands issue the KFRB09302-E message and terminate in an error).

Displaying extraction information queue file information

This command enables you to display information about the extraction information queue files at the standard output. For details about the display format, see the hdemodq command in Chapter 7. Command Syntax.

Registering additional extraction information queue files

This command enables you to initialize as much space as needed for an extraction information queue file and register the file in Datareplicator. This additional registration is not applicable when a shortage of disk space (Disk Full) is caused by write operations on an extraction information queue file. You can add a maximum of 16 extraction information queue files.

The types of extraction information queue files that can be added are UNIX regular files and character special files (and Windows files). Extraction information queue files cannot be added to the Datareplicator file system area.

An added extraction information queue file must be the same size as the existing extraction information queue files.

Releasing registered extraction information queue files

This command enables you to release extraction information queue files that have been registered in Datareplicator. This command function is applicable only to those extraction information queue files whose untransmitted extraction information status is e. The hdemodq command only releases a file's registration; it does not actually delete the extraction information queue file.

You can reduce the number of unreleased extraction information queue files to a minimum of 2. When there are only two extraction information queue files registered in Datareplicator, you can no longer perform extraction information queue file registration release.

The types of extraction information queue files whose registration can be released are UNIX regular files and character special files (and Windows files), as well as extraction information queue files in the Datareplicator file system area.

Note:

Note the following points about using the hdemodq command to change the organization of extraction information queue files:

• Backing up the extraction server status file
  The hdemodq command updates the extraction server status file because it changes the status of extraction information queue files. To be prepared for the possibility of command execution errors, always back up the extraction server status file before executing the command. If your extraction server status file is stored in the Datareplicator file system area, back up the entire Datareplicator file system area. For details about how to back up the extraction server status file, see 6.4.2(7)(b) Backing up the extraction server status file.
  
• Initialization after an organization change
  If you initialize the environment of the source Datareplicator by executing the hdestart -i command after changing the organization of the extraction information queue files with the hdemodq command, the environment is also re-constructed according to the extraction environment definition. Therefore, the system ignores the change made to the organization of the extraction information queue files by the hdemodq command.
  

The following explains how to handle the source Datareplicator's status files (extraction master status file and extraction server status file).

(a) Initializing an extraction status file

To initialize an extraction status file (extraction master status file or extraction server status file)

When using a UNIX regular file or Windows file:

1. Terminate the source system.
2. Terminate the source Datareplicator normally.
3. Execute initial start on the source Datareplicator.
4. Start the source Datareplicator normally.

Start the source system.

When using a UNIX character special file:

1. Terminate the source system.
2. Terminate the source Datareplicator normally.
3. Use an OS command to delete the previous status file.
   Perform this step only when a status file size or the disk arrangement has changed for a reason such as the addition or removal of HiRDB servers.
   
4. Use an OS command to re-create the status file in character special file format.
   Perform this step only when a status file size or the disk arrangement has changed for a reason such as the addition or removal of HiRDB servers.
   
5. Execute initial start on the source Datareplicator.
6. Start the source Datareplicator normally.
7. Start the source system.

(b) Backing up the extraction server status file

Back up the extraction server status file in the following cases:

• You will be using the hdemodq command to change the organization of extraction information queue files.
• The extraction information queue file has become full.
• You will be using the facility for recovering the extraction information queue file.

The following table shows the commands used to back up the extraction server status file.

Table 6-4 Commands used to back up the extraction server status file

Type of extraction server status file	Command to be used	Example of command execution
UNIX regular file	cp	cp $HDEPATH/sts_sds01 sts_sds01_backup
UNIX character special file	dd	dd if="$HDEPATH/sts_sds01" of=sts_sds01_backup bs=1024 count=number-of-kilobytes-in-sts_sds01
Windows file	copy	copy "%HDEPATH%\sts_sds01" sts_sds01_backup

Note 1:

In this example, the name of the extraction server status file is sts_sds01. For details about each command, see the applicable OS documentation.

Note 2:

To recover the extraction server status file from its backup, execute the command with the extraction server status file name and the backup file name swapped.

The following explains how to handle the extraction error information files (extraction master error information files and extraction node master error information files).

(a) Changing the maximum size

To change the maximum size of the extraction error information files:

1. Terminate the source system.
2. Terminate the source Datareplicator normally.
3. Use a text editor to modify the appropriate operand in the extraction system definition.
4. Start the source Datareplicator normally.
5. Start the source system.

(b) Saving an extraction error information file

When an error information file subject to data accumulation becomes full, file swapping occurs. Datareplicator re-creates the next error information file so that it can be swapped in.

To save the contents of an error information file before it is re-created:

1. Use an OS command to check the most recent update dates and times of the error information files.
2. Use an OS command to copy the error information file with the earlier update date and time into a new file under any name.

Datareplicator issues a message (KFRB00051-I or KFRB00052-I) whenever error information files are to be swapped or whenever an error information file is to be closed during Datareplicator operation; when such a message is issued, make a backup copy if necessary.

(c) Outputting error information to the standard output

When the error information file being used for data accumulation becomes full, file swapping occurs. Datareplicator re-creates the next error information file that is to be swapped in.

To output the contents of the error information file to the standard output before the file is re-created:

1. Use an OS command to check the most recent update dates and times of the error information files.
2. Use an OS command to output the extraction master error information file with the older update date and time to the standard output.

(d) Examples of error information file output

Figure 6-8 shows an example of the output from an extraction master error information file, and Figure 6-9 shows an example of the output from an extraction node master error information file.

Figure 6-8 Example of output from an extraction master error information file

Figure 6-9 Example of output from an extraction node master error information file

(e) Output destinations other than files

Datareplicator outputs the contents of error information files to the syslog file so that the information can be used to prevent the types of errors reported in the error information files and achieve automatic operation.

To output the contents of error information files to the syslog file, you must set the syslogout operand to true in the extraction system definition.

The following table shows the output destination for the contents of an extraction master error information file.

Table 6-5 Output destinations for the contents of an extraction master error information file

Status of the extraction master error information file	Output destination
	syslog file	Extraction master error information file
Normal (output enabled)	If true	Y
Error (output disabled)	If true	--

Y: Output.

If true: Output if true is specified in the syslogout operand in the extraction system definition.

--: Not output.

You can use the hdstrcedit command to view and edit the activity trace files (extraction master trace files and extraction node master trace files). For details about how to use the hdstrcedit command, see the hdstrcedit command in Chapter 7. Command Syntax.

The following explains how to handle the data linkage file.

Initializing the data linkage file

When using a UNIX regular file or Windows file:

1. Terminate the source system.
2. Terminate the source Datareplicator normally.
3. Execute initial start on the source Datareplicator.
4. Start the source Datareplicator normally.
5. Start the source system.

When using a UNIX character special file:

1. Terminate the source system.
2. Terminate the source Datareplicator normally.
3. Use an OS command to delete the previous data linkage file.
4. Use an OS command to re-create the data linkage file in character special file format.
5. Execute initial start on the source Datareplicator.
6. Start the source Datareplicator normally.
7. Start the source system.

The following explains how to handle a command log file.

The command log files contain a record of when Datareplicator commands were executed. The command log files are created automatically when Datareplicator starts. You can view a command log file at any time to check the execution history of commands. For details about the information that is output to the command log file, see Overview of commands in 7. Command Syntax.

The following figure shows an example of the contents of a command log file.

Figure 6-10 Example of command log file contents